U.S. patent application number 12/893447 was filed with the patent office on 2011-09-29 for computer-aided diagnostic systems and methods for determining skin compositions based on traditional chinese medicinal (tcm) principles.
Invention is credited to Jing Cheng, Denise M. DiCanio, Julie A. Hidalgo, Arlette Palo, Rose Marie Sparacio, Eric G. Yovine.
Application Number | 20110238604 12/893447 |
Document ID | / |
Family ID | 43857342 |
Filed Date | 2011-09-29 |
United States Patent
Application |
20110238604 |
Kind Code |
A1 |
DiCanio; Denise M. ; et
al. |
September 29, 2011 |
COMPUTER-AIDED DIAGNOSTIC SYSTEMS AND METHODS FOR DETERMINING SKIN
COMPOSITIONS BASED ON TRADITIONAL CHINESE MEDICINAL (TCM)
PRINCIPLES
Abstract
Computer-aided systems and methods are provided for determining
the skin composition of a specific user according to Traditional
Chinese Medicinal (TCM) principles by statistically analyzing
biological and/or psychological information collected from such
user, such as age, gender, bodily sensation, skin condition and
complexion, sleep pattern, dietary habits, energy level, stress
level, physical fitness and emotional wellness, so as to classify
the skin composition of the user according to TCM principles but
without employing a TCM practitioner. Preferably, the skin
composition classification is indicative of Yin-Yang balance of the
skin of the user or the lack thereof. The present systems and
methods may further recommend to the user one or more topical skin
care regimens and/or ingestible skin benefit products suitable for
the skin composition of the specific user.
Inventors: |
DiCanio; Denise M.;
(Centereach, NY) ; Palo; Arlette; (New York,
NY) ; Cheng; Jing; (Shanghai, CN) ; Sparacio;
Rose Marie; (Manorville, NY) ; Hidalgo; Julie A.;
(West Islip, NY) ; Yovine; Eric G.; (Bronx,
NY) |
Family ID: |
43857342 |
Appl. No.: |
12/893447 |
Filed: |
September 29, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61248503 |
Oct 5, 2009 |
|
|
|
Current U.S.
Class: |
706/12 ;
706/52 |
Current CPC
Class: |
G16H 10/20 20180101;
G16H 50/20 20180101 |
Class at
Publication: |
706/12 ;
706/52 |
International
Class: |
G06N 7/02 20060101
G06N007/02; G06F 15/18 20060101 G06F015/18 |
Claims
1. A computer-aided diagnosis system for determining the skin
composition of a specific user according to Traditional Chinese
Medicinal (TCM) principles, said system comprising: (1) a data
entry module configured for entering into said system a clinical
data set obtained from the specific user, wherein the clinical data
set comprises biological and/or psychological information selected
from the group consisting of age, gender, bodily sensation, tongue
color, tongue fur color, lip color, skin condition and complexion,
sleep pattern, dietary habits, energy level, stress level, physical
fitness and emotional wellness of the specific user; (2) a data
processing module in communication with the data entry module,
wherein said data processing module is configured for applying a
statistical classification process to said clinical data set so as
to classify the skin composition of the specific user according to
TCM principles; and (3) an output module in communication with the
data processing module and configured for displaying the skin
composition of the specific user as determined by the data
processing module.
2. The computer-aided diagnosis system of claim 1, wherein the skin
composition classification of the specific user is indicative of
Yin-Yang balance of the skin of the specific user or the lack
thereof.
3. The computer-aided diagnosis system of claim 1, wherein the skin
composition of the specific user is one of three different
classifications or one of five different classifications.
4. The computer-aided diagnosis system of claim 1, wherein the data
processing module is further configured for recommending one or
more topical skin care regiments and/or ingestible skin benefit
products suitable for the skin composition of the specific
user.
5. The computer-aided diagnosis system of claim 1, wherein at least
some of the biological and/or psychological information in the
clinical data set is obtained through the specific user's
self-evaluation as prompted by a questionnaire containing questions
related to one or more topics selected from the group consisting of
age, gender, bodily sensation, skin condition and complexion, sleep
pattern, dietary habits, energy level, stress level, physical
fitness, and emotional wellness of the specific user.
6. The computer-aided diagnostic system of claim 5, wherein the
self-evaluation is conducted by the specific user remotely by
filling out the questionnaire online.
7. The computer-aided diagnostic system of claim 5, wherein the
self-evaluation is conducted by the specific user on site by
filling out the questionnaire at a retail location.
8. The computer-aided diagnostic system of claim 5, wherein some of
the biological and/or psychological information in the clinical
data set is obtained through evaluation of the specific user by a
trained holistic beauty specialist.
9. The computer-aided diagnostic system of claim 8, wherein the
evaluation is conducted by the trained holistic beauty specialist
through interactive consultation with the specific user.
10. The computer-aided diagnostic system of claim 8, wherein the
evaluation is conducted on site by the trained holistic beauty
specialist at a retail location.
11. The computer-aided diagnostic system of claim 5, wherein some
of biological and/or psychological information in the clinical data
set is obtained through measurements conducted on the specific user
by an electronic device which measures one or more skin
parameters.
12. The computer-aided diagnosis system of claim 1, wherein the
statistical classification process employs a classification
technique selected from the group consisting of discriminant
analysis, logistic regression, naive Bayes classifier, support
vector machines, quadratic classifiers, neural networks,
perceptrons, decision trees, Bayesian networks, hidden Markov
models, and combinations thereof.
13. The computer-aided diagnosis system of claim 1, wherein the
statistical classification process comprises multiple
classification functions, each of which is associated with a
specific skin composition and is used to compute a classification
score for a user representing the likelihood of said user being
classified by a TCM practitioner as having the specific skin
composition associated with the respective classification function,
wherein the skin composition associated with the classification
function that renders the highest classification score for the user
is classified as the skin composition of the user.
14. The computer-aided diagnosis system of claim 13, wherein the
classification functions are either linear or quadratic.
15. The computer-aided diagnosis system of claim 13, wherein the
classification functions are constructed by a discriminant analysis
comprising: (a) obtaining a training data set that comprises
biological and/or psychological information of multiple users in a
training sample selected from a population and classification of
said users by a TCM practitioner with respect to their skin
compositions according to TCM principles; (b) identify a group of
independent variables, each of which is representative of one type
of biological and/or psychological information of the users in the
training sample; (c) calculating an F ratio for each one of the
group of independent variables, which is indicative of the
capability of said independent variable in discriminating between
different groups of users of different skin compositions as
determined by the TCM practitioner, wherein the F ratio for a
specific independent variable is calculated as the ratio of
between-group variance of said independent variable over the
average within-group variance of said independent variable; (d)
selecting the independent variable with the largest F ratio; (e)
calculating F ratios for the remaining independent variables; (f)
repeating steps (d)-(e) until all independent variables having F
ratios greater than a pre-set level of significance have been
selected; and (g) using the selected independent variables to
construct the multiple classification functions.
16. The computer-aided diagnosis system of claim 15, wherein each
classification function is a linear function having a general
formula C.sub.x+W.sub.x1.times.V.sub.1+W.sub.x2.times.V.sub.2+ . .
. +W.sub.xm.times.V.sub.m, wherein C.sub.x is a constant associated
with a specific skin composition, V.sub.1 to V.sub.m are the
selected independent variables, W.sub.x1 to W.sub.xm are
coefficients that each corresponds with one of the selected
independent variables for the skin composition and are indicative
of the relative weights of the selected variables in computing the
classification score using the respective classification
function.
17. The computer-aided diagnosis system of claim 15, wherein the
classification functions are validated by using a validation data
set that comprises biological and/or psychological information of
multiple users in a validation sample and classification of said
users by a TCM practitioner with respect to their skin
compositions, and wherein said validation sample is selected from
the same population as the training sample but contain a different
group of users.
18. A computer-readable medium comprising code for determining the
skin composition of a specific user according to Traditional
Chinese Medicinal (TCM) principles, said code comprising
instruction for applying a statistical classification process to a
clinical data set obtained from the specific user so as to classify
the skin composition of said specific user according to TCM
principles, wherein said clinical data set comprises biological
and/or psychological information selected from the group consisting
of age, gender, bodily sensation, tongue color, tongue fur color,
lip color, skin condition and complexion, sleep pattern, dietary
habits, energy level, stress level, physical fitness and emotional
wellness of the specific user.
19. The computer-readable medium of claim 18, wherein the skin
composition classification of the specific user is indicative of
Yin-Yang balance of the skin of the user or the lack thereof.
20. The computer-readable medium of claim 18, wherein the skin
composition of the specific user is selected from one of three
different classifications or one of five different
classifications.
21. The computer-readable medium of claim 18, further comprising
instruction for recommending one or more topical skin care
regiments and/or ingestible skin benefit products suitable for the
skin composition of the specific user.
22. The computer-readable medium of claim 18, wherein the
statistical classification process comprises multiple
classification functions, each of which is associated with a
specific skin composition and is used to compute a classification
score for a user representing the likelihood of said user being
classified by a TCM practitioner as having the specific skin
composition associated with the respective classification function,
wherein the skin composition associated with the classification
function that renders the highest classification score for the user
is classified as the skin composition of the user.
23. The computer-readable medium of claim 22, wherein each
classification function is a linear function having a general
formula C.sub.x+W.sub.x1.times.V.sub.1+W.sub.x2.times.V.sub.2+ . .
. +W.sub.xm.times.V.sub.m, wherein C.sub.x is a constant associated
with a specific skin composition, V.sub.1 to V.sub.m are selected
independent variables that each represents one type of biological
and/or psychological information of the user, W.sub.x1 to W.sub.xm
are coefficients each corresponding with one of the selected
independent variables for the skin composition and are indicative
of the relative weights of the selected independent variables in
computing the classification score using the respective
classification function.
24. A method for determining the skin composition of a specific
user according to Traditional Chinese Medicinal (TCM) principles,
comprising: (1) obtaining a clinical data set from the specific
user, which comprises biological and/or psychological information
selected from the group consisting of age, gender, bodily
sensation, tongue color, tongue fur color, lip color, skin
condition and complexion, sleep pattern, dietary habits, energy
level, stress level, physical fitness and emotional wellness of the
specific user; (2) applying a statistical classification process to
said clinical data set of the specific user for classifying the
skin composition of the specific user according to TCM principles;
and (3) outputting the skin composition of the specific user so
determined.
25. The method of claim 24, wherein the skin composition
classification of the specific user is indicative of Yin-Yang
balance of the skin of the user or the lack thereof.
26. The method of claim 24, wherein the skin composition of the
specific user is selected from one of three different
classifications or one of five different classifications.
27. The method of claim 24, further comprising recommending one or
more topical skin care regiments and/or ingestible skin benefit
products suitable for the skin composition of the specific
user.
28. The method of claim 24, wherein at least some of the biological
and/or psychological information in the clinical data set is
obtained through the specific user's self-evaluation as prompted by
a questionnaire containing questions related to one or more topics
selected from the group consisting of age, gender, bodily
sensation, skin condition and complexion, sleep pattern, dietary
habits, energy level, stress level, physical fitness, and emotional
wellness of the specific user.
29. The method of claim 28, wherein the self-evaluation is
conducted by the specific user remotely by filling out the
questionnaire online.
30. The method of claim 28, wherein the self-evaluation is
conducted by the specific user on site by filling out the
questionnaire at a retail location.
31. The method of claim 28, wherein some of the biological and/or
psychological information in the clinical data set is obtained
through evaluation of the specific user by a trained holistic
beauty specialist.
32. The method of claim 31, wherein the evaluation is conducted by
the trained holistic beauty specialist through interactive
consultation with the specific user.
33. The method of claim 31, wherein the evaluation is conducted on
site by the trained holistic beauty specialist at a retail
location.
34. The method of claim 28, wherein some of biological and/or
psychological information in the clinical data set is obtained
through measurements conducted on the specific user by an
electronic device.
35. The method of claim 24, wherein the statistical classification
process employs a classification technique selected from the group
consisting of discriminant analysis, logistic regression, naive
Bayes classifier, support vector machines, quadratic classifiers,
neural networks, perceptrons, decision trees, Bayesian networks,
hidden Markov models, and combinations thereof.
36. The method of claim 24, wherein the statistical classification
process comprises multiple classification functions, each of which
is associated with a specific skin composition and is used to
compute a classification score for a user representing the
likelihood of said user being classified by a TCM practitioner as
having the specific skin composition associated with the respective
classification function, wherein the skin composition associated
with the classification function that renders the highest
classification score for the user is classified as the skin
composition of the user.
37. The method of claim 36, wherein the classification functions
are either linear or quadratic.
38. The method of claim 36, wherein the classification functions
are constructed by a discriminant analysis comprising: (a)
obtaining a training data set that comprises biological and/or
psychological information of multiple users in a training sample
selected from a population and classification of said users by a
TCM practitioner with respect to their skin compositions according
to TCM principles; (b) identifying a group of independent
variables, each of which is representative of one type of
biological and/or psychological information of the users in the
training sample; (c) calculating an F ratio for each one of the
group of independent variables, which is indicative of the
capability of said independent variable in discriminating between
different groups of users of different skin compositions as
determined by the TCM practitioner, wherein the F ratio for a
specific independent variable is calculated as the ratio of
between-group variance of said independent variable over the
average within-group variance of said independent variable; (d)
selecting the independent variable with the largest F ratio; (e)
calculating F ratios for the remaining independent variables; (f)
repeating steps (d)-(e) until all independent variables having F
ratios greater than a pre-set level of significance have been
selected; and (g) using the selected independent variables to
construct the multiple classification functions.
39. The method of claim 38, wherein each classification function is
a linear function having a general formula
C.sub.x+W.sub.x1.times.V.sub.1+W.sub.x2.times.V.sub.2+ . . .
+W.sub.xm.times.V.sub.m, wherein C.sub.x is a constant associated
with a specific skin composition, V.sub.1 to V.sub.m are the
selected independent variables, W.sub.x1 to W.sub.xm are
coefficients that each corresponds with one of the selected
independent variables for the skin composition and are indicative
of the relative weights of the selected variables in computing the
classification score using the respective classification
function.
40. The method of claim 38, wherein the classification functions
are validated by using a validation data set that comprises
biological and/or psychological information of multiple users in a
validation sample and classification of said users by a TCM
practitioner with respect to their skin compositions, and wherein
said validation sample is selected from the same population as the
training sample but contain a different group of users.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority from U.S. Provisional
Patent Application No. 61/248,503, filed Oct. 5, 2009.
FIELD OF THE INVENTION
[0002] The present invention relates to computer-aided diagnostic
systems and methods for determining the skin composition of a
specific user according to Traditional Chinese Medicinal (TCM)
principles.
BACKGROUND OF THE INVENTION
[0003] There is great interest in the cosmetic industry to develop
products containing natural plant-derived materials that may be
applied topically to the skin to provide anti-acne, anti-oil, and
anti-cellulite benefits. Plant-based cosmetic products that enhance
the appearance of skin are increasingly in demand. Active
ingredients or components with skin care benefits can be obtained
from either the entire plant or various parts of a plant, such as
seeds, needles, leaves, roots, bark, cones, stems, rhizomes, callus
cells, protoplasts, organs and organ systems, and meristems, either
as dried powders or liquid extracts, which can then be incorporated
into topical compositions.
[0004] Traditional Chinese Medicine (hereinafter "TCM") has been in
existence for several thousands of years and is based largely on
accumulated human experience in using naturally occurring plant and
animal extracts to treat various diseases. In recently years, TCM
practices have gained significant recognition by the Western world,
especially for treatment of chronic pathological conditions.
Certain TCM ingredients have been known to have skin care benefits
for thousands of years, and cosmetic or skin care products
containing such TCM ingredients are becoming more popular. Because
the TCM ingredients are all natural, they are significantly more
appealing to customers who are conscious of the side effects and
adverse environmental impacts of synthetic compounds.
[0005] However, it is important to note that the basic principles
of TCM are quite different from that of the western medicinal
sciences. For example, TCM takes a holistic view of the human
bodies and believes that an internal physiological balance has to
be maintained for a person to be healthy, and all diseases are
caused by one or more types of imbalances, and each type of
imbalances manifests through an identifiable pattern of symptom and
requires a particular set of treatments for correction. Typically,
a TCM practitioner first exams an individual patient through
various non-invasive techniques, including observing certain
physiological characteristics of such patient, listening to the
patient's voices, sensing any change in the patient's bodily odor
or temperature, feeling the patient's pulse, asking various
questions about the history and development of the disease, and
gathering any other information that may be related to the
patient's health condition, before rendering an opinion on the
specific type of imbalance or disease the patient is suffering.
Then the TCM practitioner will prescribe a blend of multiple herbal
ingredients for treating various symptoms of the disease and
helping to re-establish the internal physiological balance of the
patient. In other words, the TCM practice takes an individualized
treatment approach, instead of a "one-size-fits-all" approach.
[0006] The currently available skin care products containing
naturally occurring TCM ingredients are mass-marketed to all
consumers and fail to implement the individualized treatment
approach of TCM. Therefore, it would be desirable to provide new
cosmetic and skin care products containing TCM ingredients that
adopt the individualized treatment approach of TCM in improving the
appearance and skin condition of specific users.
[0007] However, a major challenge for implementing the
individualized treatment approach of TCM is that a correct
diagnosis requires proper application of very complicated TCM
principles, which were established through trial-and-error over
thousands of years and which are still not well understood by the
Western world to this date. Only very experienced TCM practitioners
can master such an art of diagnosis, but it is cost-prohibitive to
employ a TCM practitioner at each retail location for diagnosing
patients and recommending skin care products appropriate for the
individual patients.
[0008] Therefore, it is an object of the present invention to
provide computer-aided diagnostic systems and methods for rapidly
and accurately determining the skin compositions of a specific user
according to TCM principles and optionally for recommending skin
care products suitable for the skin composition of the specific
user, but without having to employ a TCM practitioner at each
retail location.
SUMMARY OF THE INVENTION
[0009] The present invention in one aspect relates to a
computer-aided diagnostic system for determining the skin
composition of a specific user according to Traditional Chinese
Medicinal (TCM) principles, said system comprising: [0010] (1) a
data entry module configured for entering into the system a
clinical data set obtained from the specific user, wherein the
clinical data set contains biological and/or psychological
information selected from the group consisting of age, gender,
bodily sensation, tongue color, tongue fur color, lip color, skin
condition and complexion, sleep pattern, dietary habits, energy
level, stress level, physical fitness and emotional wellness of the
specific user; [0011] (2) a data processing module in communication
with the data entry module, wherein the data processing module is
configured for applying a statistical classification process to the
clinical data set so as to classify the skin composition of the
specific user according to TCM principles; and [0012] (3) an output
module in communication with the data processing module and
configured for displaying the skin composition of the specific user
as determined by the data processing module.
[0013] Preferably, the skin composition classification of the
specific user is indicative of Yin-Yang balance of the skin of the
specific user or the lack thereof. The data processing module may
be further configured for recommending one or more topical skin
care regiments and/or ingestible skin benefit products for the skin
composition of the specific user.
[0014] In another aspect, the present invention is related to a
computer-readable medium comprising code for determining the skin
composition of a specific user according to Traditional Chinese
Medicinal (TCM) principles, such code comprising instruction for
applying a statistical classification process to a clinical data
set obtained from the specific user so as to classify the skin
composition of the specific user according to TCM principles,
wherein the clinical data set comprises biological and/or
psychological information selected from the group consisting of
age, gender, bodily sensation, tongue color, tongue fur color, lip
color, skin condition and complexion, sleep pattern, dietary
habits, energy level, stress level, physical fitness and emotional
wellness of the specific user.
[0015] In yet another aspect, the present invention is related to a
method for determining the skin composition of a specific user
according to Traditional Chinese Medicinal (TCM) principles,
comprising: [0016] (1) obtaining a clinical data set from the
specific user, which comprises biological and/or psychological
information selected from the group consisting of age, gender,
bodily sensation, tongue color, tongue fur color, lip color, skin
condition and complexion, sleep pattern, dietary habits, energy
level, stress level, physical fitness and emotional wellness of the
specific user; [0017] (2) applying a statistical classification
process to said clinical data set of the specific user for
classifying the skin composition of the specific user according to
TCM principles; and [0018] (3) outputting the skin composition of
the specific user so determined.
[0019] The above-mentioned biological and/or psychological
information of the user can be collected through different sources.
For example, at least some of the biological and/or psychological
information in the clinic data set is obtained through the specific
user's self-evaluation as prompted by a questionnaire containing
various questions related to one or more topics selected from the
group consisting of age, gender, bodily sensation, skin condition
and complexion, sleep pattern, dietary habits, energy level, stress
level, physical fitness, and emotional wellness of the user. Such
self-evaluation can be conducted either remotely by allowing the
user to fill out the questionnaire online, or on site at a retail
location. Some of the biological and/or psychological information
can also be obtained through evaluation of the specific user by a
trained holistic beauty specialist, and such evaluation can be
conducted either remotely through interactive consultation via the
internet or on site at a retail location. Further, some of the
biological and/or psychological information can be obtained through
measurements conducted on the specific user by an electronic
device.
[0020] The statistical classification process as employed by the
present invention may employ any suitable classification technique,
such as discriminant analysis, logistic regression, naive Bayes
classifier, support vector machines, quadratic classifiers, neural
networks, perceptrons, decision trees, Bayesian networks, hidden
Markov models, and combinations thereof. The statistical
classification process may comprise multiple classification
functions, each of which is associated with a specific skin
composition and is used to compute a classification score for a
user representing the likelihood of the user being classified by a
TCM practitioner as having the specific skin composition associated
with the respective classification function, wherein the skin
composition associated with the classification function that
renders the highest classification score for the user is classified
as the skin composition of the user. The classification functions
can be either linear or quadratic functions.
[0021] In a preferred, but not necessary, embodiment of the present
invention, the classification functions are constructed by a
discriminant analysis comprising: [0022] (a) obtaining a training
data set that comprises biological and/or psychological information
of multiple users in a training sample selected from a population
and classification of such users by a TCM practitioner with respect
to their skin compositions according to TCM principles; [0023] (b)
identify a group of independent variables, each of which is
representative of one type of biological and/or psychological
information of the users in the training sample; [0024] (c)
calculating an F ratio for each one of the group of independent
variables, which is indicative of the capability of the independent
variable in discriminating between different groups of users of
different skin compositions as determined by the TCM practitioner,
wherein the F ratio for a specific independent variable is
calculated as the ratio of between-group variance of such
independent variable over the average within-group variance of such
independent variable; [0025] (d) selecting the independent variable
with the largest F ratio; [0026] (e) calculating F ratios for the
remaining independent variables; [0027] (f) repeating steps (d)-(e)
until all independent variables having F ratios greater than a
pre-set level of significance have been selected; and [0028] (g)
using the selected independent variables to construct the multiple
classification functions.
[0029] In a more preferred embodiment of the present invention,
each of the classification functions is a linear function having a
general formula
C.sub.x+W.sub.x1.times.V.sub.1+W.sub.x2.times.V.sub.2+ . . .
+W.sub.xm.times.V.sub.m, wherein C.sub.x is a constant associated
with a specific skin composition V.sub.1 to V.sub.m are the
selected independent variables, W.sub.x1 to W.sub.xm are
coefficients, each of which corresponds with one of the selected
independent variables for the skin composition, and each of which
is indicative of the relative weights of the selected variables in
computing the classification score using the respective
classification function.
[0030] Other aspects and objectives of the present invention will
become more apparent from the ensuring description, examples, and
claims.
DETAILED DESCRIPTION OF THE INVENTION, AND PREFERRED EMBODIMENT
THEREOF
Introduction
[0031] In TCM practices, health is represented as a balance of Yin
and Yang, which are opposing forces interconnected and
interdependent in the natural world, giving rise to each other in
turn. These two forces represent the bipolar manifestation of all
things in nature, and because of this, one must be present to allow
the other to exist. A constantly changing balance between Yin and
Yang should be maintained within the body of a person. When one
force is regularly dominating over the other, the health of the
patient can become compromised, resulting in illness and disease.
Therefore, when diagnosing a patient, TCM practitioners usually try
to first determine the exact nature of the Yin/Yang imbalance, and
then correct it through the use of acupuncture, herbal remedies,
exercise, diet and lifestyle. Once the balance is restored in the
body, so is health.
[0032] Therefore, according to the TCM principles, the skin
composition of a patient can be classified according to the
Yin-Yang balance or imbalance in the skin of such user. For
example, the skin composition of the user may be classified into
one of three categories, including Yang-dominant, balanced, and
Yin-dominant. For another example, the skin composition of the user
may be classified into one of five different classifications,
including Yang-dominant, balanced-to-Yang, balanced,
balanced-to-Yin, and Yin-dominant.
[0033] An experienced TCM practitioner typically determines the
Yin-Yang balance or imbalance of a patient by observing various
physiological characteristics of such patient, such as tongue
color, tongue fur color, lip color, skin condition and complexion,
and the like. Further, the TCM practitioner asks various questions
regarding the patient's bodily sensation, sleep pattern, dietary
habits, energy level, stress level, physical fitness and emotional
wellness, the answers to which may further assist the TCM
practitioner in determining the exact type of Yin-Yang balance or
imbalance that the patient is experiencing. The TCM practitioner
may also listen to the patience's voices, sensing any change in the
patient's bodily odor or temperature, feeling the patient's pulse,
and gather any other information related to the patient's health
condition through non-invasive techniques, which will be taken into
consideration for diagnosis.
[0034] For example, the balanced skin composition is typically
characterized by smooth, radiant, soft and delicate skin with good
elasticity and even skin tone, and patients with the balanced skin
composition usually also have lips and tongue of light red color
with thin, pale white tongue fur color. The Yang-dominant skin
composition is typically characterized by reddish cheek and red
skin tone, oily skin surface, and tendency to develop allergy,
acne, pigmentation, and pre-mature wrinkles. Patients with the
Yang-dominant skin composition may also have lips and tongue of
bright red color with yellow tongue fur color, and they typically
prefer cold drinks, become thirsty very easily, have bitter tastes
in the mouth, often sweat when sleeping at night, easily irascible,
and often suffer from internal heat. In contrast, the Yin-dominant
skin composition is typically characterized by pale skin tone, dry
and easily desquamated skin with low water absorption and water
retention capacity, and tendency to develop pigmentation and
pre-mature wrinkles. Patients with the Yin-dominant skin
composition may have pale or blue lips, light red or pale tongue
color with white tongue fur color, and they typically prefer hot
drinks, have relatively cold hands and feet, are sensitive to
temperature decrease, and often feel tired or sleepy.
[0035] Determination of the Yin-Yang balance or imbalance of a user
according to TCM principles is a very complex task, and only
experienced TCM practitioners can make an accurate and speedy
diagnosis. Therefore, it presents a challenge for the cosmetic
industry to provide TCM-based cosmetic and skin care products for
specific users based on his/her particular skin composition to help
either maintaining or restoring the Yin-Yang balance in his/her
skin.
[0036] The present invention is based, in part, upon the surprising
discovery that a computer-aided system and method can be used to
statistically model diagnoses rendered by an experienced TCM
practitioners for a group of users from a training sample with
respect to their skin compositions, based on certain biological
and/or psychological information learned of such users, and
subsequently to predict or classify the unknown skin composition of
a user based on biological and/or psychological information learned
of such user. In some aspects, the present invention uses
statistical algorithms to classify a user's skin composition into a
particular category according to TCM principles. In other aspects,
the present invention uses statistical algorithms for ruling out
the possibility that a user's skin composition belongs to a
particular category to aid in the classification of the user's skin
composition into a different category.
DEFINITIONS
[0037] As used herein, the following terms have the meanings as
ascribed to them as follows:
[0038] The term "classifying" includes "to predict" or "to
categorize" a sample with a skin composition. In certain instances,
"classifying" is based on statistical evidence, empirical evidence,
or both. In certain embodiments, the systems and methods of the
present invention use a so-called training sample with known skin
compositions as diagnosed by an experienced TCM practitioner
according to TCM principles, and collect certain biological and/or
psychological information that may be used by the TCM practitioner
in rendering the diagnoses from the training sample. Once
established, the training data set serves as a basis, model, or
template against which the biological information of a user with
unknown skin composition is compared, in order to predict into
which category the user's skin composition is most likely to be
classified by the TCM practitioner according to TCM principles. In
certain instances, classifying a particular user is akin to
determine the skin composition of the user. In certain other
instances, classifying the user is akin to differentiate the skin
composition of the user from another type of skin composition.
[0039] The term "skin composition" as used herein refers to various
categories of a user's skin based on the Yin-Yang balance or
imbalance of such user according to TCM principles. In the two
specific embodiments of the present invention, the skin
compositions of a population are either classified into three
categories, including Yang-dominant, balanced, and Yin-dominant, or
five categories, including Yang-dominant, balanced-to-Yang,
balanced, balanced-to-Yin, and Yin-dominant. However, it is
important to note that the actual number of categories is not
limited to the specific embodiments illustrated herein.
[0040] The term "biological information" as used herein refers to
any set of physiological or biological data that represents the
distinctive features or characteristics of a user associated with a
particular skin composition. For example, the biological and/or
psychological information of a user can include age, gender, bodily
sensation, tongue color, tongue fur color, lip color, skin
condition and complexion, sleep pattern, dietary habits, energy
level, stress level, physical fitness and emotional wellness of the
user.
Description of the Embodiments
[0041] The present invention provides systems and methods for
accurately classifying the skin composition of a user according to
TCM principles. The present invention implements a statistical
classification process to predict the skin compositions of new
users according to TCM principles, but without employing a TCM
practitioner. The statistical classification process of the present
invention may be based on one or more widely used classification
techniques, and it uses data collected from certain users in a
training sample with known skin compositions (i.e., as diagnosed by
an experienced TCM practitioner) to construct one or more
classification functions that best correlate the biological and/or
psychological information or data of users in the training sample
with their skin compositions as diagnosed by the TCM practitioner.
Such classification functions can then be used to predict the skin
compositions of new users based solely on the biological and/or
psychological information or data of such new users, without
involving the TCM practitioner.
Biological Information Collection
[0042] In order to statistically model the diagnoses of an
experienced TCM practitioner, the present invention collects
various biological and/or psychological information of the user
that may be useful in assisting the TCM practitioner in determining
the skin composition of such user. For example, such biological
and/or psychological information may include, but is not limited
to: age, gender, bodily sensation, tongue color, tongue fur color,
lip color, skin condition and complexion, sleep pattern, dietary
habits, energy level, stress level, physical fitness and emotional
wellness of the user.
[0043] At least some of the biological and/or psychological
information can be collected by asking the specific user to conduct
certain self-evaluation and fill out a questionnaire containing
questions related to the above-listed topics. The questionnaire can
be completed by the specific user on site at a retain location, or
it can be completed remotely online. In some embodiments, the
questionnaire comprises a first section containing a set of
questions asking the specific user to provide answers regarding
certain clinical symptoms typically taken into consideration by a
TCM practitioner when examining a patient. For example, the
questions may be directed to certain bodily sensations experienced
by the specific user, such as feeling of heat at one's palms,
tendency of one to sweat (either in general or at certain specific
locations, such as hands or feet), tendency of one to feel thirsty
and want to drink, tendency of one to feel tired and sleep,
tendency of one to suffer from internal heat (which may be
manifested as hot flushes or oral ulcers), tendency of one to taste
bitter in one's mouth, energy level felt by oneself, and the like.
The questions may also be directed to certain observable
physiological symptoms of the specific users, such as the color of
one's tongue, tongue fur, or lips, the overall skin complexion, and
the like. The questionnaire may comprise a second section
containing a set of questions asking the specific user to provide
answers with more direct relevance to the skin problems suffered by
the user. For example, the questions may be directed to skin
dryness and/or redness, acne breakouts, pore sizes, unevenness of
skin tone, undesirable pigmentation, sebum secretion, existence of
lines and wrinkles, sagginess, skin sensitivity, and the like. The
questionnaire may comprise a third section containing a set of
questions asking the specific user to provide answers relating to
his or her life style, such as one's sleep pattern, dietary habits
(including smoking, drinking, vitamin intake, caffeine intake, and
the like), physical fitness (including exercise frequency and
intensity), stress level, emotional wellness, and the like.
Optionally, one or more trained holistic beauty specialists may
assist the user in filling out certain questions of the
questionnaire, such as those related to the tongue color, tongue
fur color, lip color, skin complexion, and skin problems of the
user. Upon completion of the questionnaire by the user, the answers
to the questions can be converted into numerical values for future
data processing.
[0044] Preferably, but not necessarily, some of the biological
and/or psychological information, especially those related to the
tongue color, tongue fur color, lip color, skin complexion, and
skin problems of the user, may be collected from by a trained
holistic beauty specialist, who examines the user and evaluates the
particular feature or characteristic of the user according to photo
grading scales, so as to improve the objectivity and accuracy of
the evaluation. The trained holistic beauty specialist can carry
out the user evaluation either on site at a retail location, or
remotely through interactive consultation with the specific user
via the internet (e.g., by using certain widely available instant
messaging application software with video-conferencing capacity,
such as iChat or Skype).
[0045] Further, some of the biological and/or psychological
information may optionally be collected by using an electronic
device that takes certain measurements on the specific user. For
example, various imaging devices can be used to quantitatively
measure coloration of certain parts of the user's body, such as the
face, cheeks, lips, tongue, tongue fur, and the like. For another
example, various medical devices can be used to quantitatively
measure the user's pulse, heart rate, blood pressure, blood glucose
level, body fat percentage, energy field, overall blood circulation
or local blood circulation, and the like. Some examples of various
electronic devices which may be used in collecting information
include a Corneometer.RTM. CM 825, measuring skin surface
hydration; a Sebumeter.RTM. SM 815, measuring sebum on the skin
surface; a Cutometer.RTM., measuring skin elasticity; a
Skin-pH-Meter.RTM. PH 905; a Mexameter.RTM. MX 18, measuring
melanin and/or erythema; a Skin-Thermometer.RTM. ST 500, measuring
skin temperature; a Tewameter.RTM. ST 500; a Colorimeter CL 400,
measuring redness; and a Glossymeter GL 490, measuring
radiance/dullness, all of which are available from Courage+Khazaka
Electronic GmbH, Germany. Such measurement results can then be
incorporated into the clinical data set of the user as potential
predictors for classifying the skin composition of the user.
Establishment of Training Data Set
[0046] In order to establish statistical models for accurately
predicting the skin compositions of new users, a training data set
must first be obtained, which contains biological and/or
psychological information of multiple users from a training sample
and the skin compositions of such users as determined by a TCM
practitioner according to TCM principles. The training sample is
typically selected from the same population from which the new
users are to be selected, and the skin compositions of the users in
the training sample, since they are already classified by a TCM
practitioner according to TCM principles, are therefore deemed as
"known." Such training data set provides training models to which
statistical classification processes or analyses can be applied in
order to formulate either linear or nonlinear classification
functions for predicting or classifying the unknown skin
composition of a new user based on his or her biological and/or
psychological information.
Statistical Classification Analysis
[0047] The term "statistical classification process" or
"statistical classification analysis" as used herein includes any
known classification technique for placing subjects into categories
based on knowledge about one or more characteristics or traits
inherent in the subjects. The inherent characteristics or traits
are typically considered independent variables and referred to as
predictors (X), and the category to which the subject is classified
is typically considered a dependent variable and referred to as the
grouping variable (Y). The goal of statistical classification
process or analysis is to predict the group variable Y for a new
subject based on a set of predictors X associated with this new
subject and a training set of previously grouped subjects. In the
present invention, the predictors are the biological and/or
psychological information of a user that can be obtained without
involvement of a TCM practitioner, and the grouping variable is the
skin composition of such user as determined by the TCM practitioner
according to TCM principles.
[0048] Preferably, but not necessarily, the statistical
classification process or analysis of the present invention employs
one or more classification techniques selected from the group
consisting of discriminant analysis, logistic regression, naive
Bayes classifier, support vector machines, quadratic classifiers,
neural networks, perceptrons, decision trees, Bayesian networks,
hidden Markov models, and combinations thereof. Performance of a
specific classification technique depends greatly on the
characteristics of the data to be classified. There is no single
classification technique that works best on all given problems.
Various empirical tests have been performed to compare performance
of classification techniques and to find the characteristics of
data that optimize the performance of classification techniques.
However, determining a suitable classification technique for a
given problem is still more of an art than a science to this
date.
[0049] In the present invention, a discriminant analysis, and more
precisely a multi-class linear discriminant analysis, is found to
be particularly effective in categorizing the skin compositions of
users according to TCM principles. Specifically, the discriminant
analysis is carried out by obtaining a training data set as
described hereinabove, i.e., which comprises biological and/or
psychological information of multiple users in a training sample
selected from a population and classification of such users by a
TCM practitioner with respect to their skin compositions according
to TCM principles. A group of independent variables or predictors
(V) is identified, each variable or predictor being representative
of one type of biological and/or psychological information of the
users in the training sample. An F ratio is calculated for each one
of the group of independent variables, which is indicative of the
capability of the independent variable in discriminating between
different groups of users of different skin compositions as
determined by the TCM practitioner. More specifically, the F ratio
for a particular independent variable is calculated as the ratio of
between-group variance of such independent variable over the
average within-group variance of such independent variable. The
independent variable with the largest F ratio is initially selected
and taken out of the pool of independent variables. The selected
independent variable will be included into the classification
functions, which are to be constructed once all independent
significant discriminating variables between different groups of
users are selected. F ratios for the remaining independent
variables (the pool excluding the already selected independent
variable) are then calculated, and the independent variable with
the largest F ratio among these is then selected and taken out of
the pool of independent variables. The calculation and selection
steps are reiterated until all independent variables having F
ratios greater than a pre-set level of significance have been
selected.
[0050] The selected independent variables are then used to
construct classification functions, which can be used to determine
which category of skin composition a new user most likely belongs.
There are as many classification functions as the number of skin
composition categories. Each classification function is associated
with a specific skin composition category and can be used to
compute a classification score for a new user by entering the
biological and/or psychological information of such new user. The
classification score outputted by each classification function
represents the likelihood of such new user being classified by a
TCM practitioner as having the specific skin composition associated
with the respective classification function. In other words, for
each new user, each classification function will compute a
classification score, and the skin composition associated with the
classification function that renders the highest classification
score is the skin composition of the user.
[0051] The classification functions can be either linear or
quadratic. In a preferred embodiment of the present invention, the
classification functions are linear having a general formula
C.sub.x+W.sub.x1.times.V.sub.1+W.sub.x2.times.V.sub.2+ . . .
+W.sub.xm.times.V.sub.m, where C.sub.x is a constant associated
with a specific skin composition, V.sub.1 to V.sub.m are the
selected independent variables, W.sub.x1 to W.sub.xm are
coefficients that each correspond with one of the selected
independent variables for each skin composition and are indicative
of the relative weights of the selected variables in computing the
classification score using the respective classification functions.
For example, assuming that the skin compositions can be divided
into three categories A, B, and C, and that five independent
variables V.sub.1, V.sub.2, V.sub.3, V.sub.4, and V.sub.5 have been
selected for constructing the classification functions, there will
be three (3) classification functions, each of which is associated
with a specific skin composition category. For a specific user with
a specific set of values for independent variables V.sub.1 through
V.sub.5, three classification scores Y.sub.A, Y.sub.B, and Y.sub.C
will be calculated as follows:
Y.sub.A=C.sub.A+W.sub.A1.times.V.sub.1+W.sub.A2.times.V.sub.2+W.sub.A3.t-
imes.V.sub.3+W.sub.A4.times.V.sub.4+W.sub.A5.times.V.sub.5,
Y.sub.B=C.sub.B+W.sub.B1.times.V.sub.1+W.sub.B2.times.V.sub.2+W.sub.B3.t-
imes.V.sub.3+W.sub.B4.times.V.sub.4+W.sub.B5.times.V.sub.5, and
Y.sub.C=C.sub.C+W.sub.C1.times.V.sub.1+W.sub.C2.times.V.sub.2+W.sub.C3.t-
imes.V.sub.3+W.sub.C4.times.V.sub.4+W.sub.C5.times.V.sub.5,
[0052] If Y.sub.A is the largest among all three classification
scores computed, it is then concluded that the specific user is
most likely to be classified as having skin composition A. However,
if Y.sub.C is the largest among all three classification scores
computed, it will be concluded that the specific user is most
likely to be classified as having skin composition C.
[0053] The discriminant analysis as described hereinabove can be
readily conducted using widely used statistical computer software
programs, such as SPSS, SAS, and the like. Optionally, the
classification functions so constructed are further validated using
a validation data set, which comprises biological and/or
psychological information of multiple users in a validation sample
and the skin compositions of such users as determined by a TCM
practitioner according to TCM principles. The validation sample is
selected from the same population as the training sample, but it
contains a different group of users. The prediction made by using
the classification functions and the actual skin composition as
determined by the TCM practitioner for the users in the validation
sample are compared, and a prediction score may be calculated
indicative of the percentage of correct prediction made by using
the classification functions. If the prediction score is higher
than a pre-set threshold, such as, for example 70%, 80%, or 90%,
the classification functions are then considered validated and
suitable for clinical uses.
[0054] The validation sample and the training sample can be
selected at the same time. Further, the validation and training
phases are not necessarily separated and fixed. The system can be
interactive, and the training and validation can be interwoven for
fine-tuning the discriminant analysis in order to achieve the best
predictions possible.
Computer-Based System
[0055] The computer-based system of the present invention may
include a computer, which can be either a workstation or a personal
computer. Such a computer preferably contains a data entry module,
such as a keyboard or a touch-screen, to allow direct entry of
users' biological and/or psychological information by the users or
by the holistic beauty specialist. The data entry module may also
be a direct connection (such as USB, Firewire, or other interface)
or a network connection (such as modem or other network
connections) for transmitting and receiving use-related biological
and/or psychological data, for example, via a portable medium such
as a compact disk (CD) or a digital versatile disk (DVD) or via the
internet.
[0056] The computer further contains a data processing module, such
as a central processing unit (CPU), which is in communication with
the data entry module. The data processing module is configured for
applying the above-described statistical classification process to
the user data entered through the data entry module and to thereby
classify the skin composition of the specific user according to TCM
principles. The data processing module is more specifically under
the control of a computer-readable medium containing coded
instructions for applying such statistical classification process
to the user data.
[0057] The computer further contains an output or displaying module
in communication with the data processing module for displaying the
result determined, i.e., the likely skin composition of the user.
Such an output or displaying module may include the computer
display screen for visually displaying the result, or a printer for
printing the result, or a network connection for outputting the
result to a remote location.
[0058] Optionally, the computer has a memory module that stores
information about various topical skin care regimens and/or
ingestible products that are suitable for different skin
compositions. Once the skin composition of a particular user is
determined by the data processing module, information about skin
care regimens and/or ingestible products suitable for the skin
composition of such particular user can be retrieved from the
memory module and output together with the skin composition
determined as skin care recommendations for the user.
[0059] In this manner, the present invention can accurately
recommend TCM-based skin care products for targeting any
sub-population of users (e.g., in any region, country, city, etc.)
with a specific skin composition according to the TCM principles,
but without employing a TCM practitioner at each retail
location.
Example 1
[0060] A group of 133 Asian female participants between the ages of
22 and 51 was recruited for testing. After registration, each
participant was given a questionnaire containing 48 questions
covering various aspects of the participant's biological and
psychological conditions. For example, the questionnaire may
inquire the participant about her age, sensory feel (e.g., fear of
cold/heat, sensation of heat at one's palms, palm sweating,
sweating pattern, feeling of thirst, feeling of bitterness in one's
mouth, suffering from internal heat, and the like), sleep pattern,
dietary habits, skin conditions (e.g., skin color/tone,
dryness/oiliness, and the like), facial complexion and tones (e.g.,
redness, darkening, and the like), skin problems (e.g., acne
breakouts, pores, uneven skin tone, hyperpigmentation, dark spots,
sensitivity, dry patches, lines/wrinkles, sagging, and the like),
energy level, stress level, physical and emotional well-being, and
the like. An exemplary questionnaire that can be used for
practicing the present invention is included herein. It is provided
for illustration purposes only, without intending to limit the
scope of the present invention in any manner.
TABLE-US-00001 TABLE 1 EXEMPLARY QUESTIONNAIRE No. Question Answers
1 What is your age? (A) 20~29 (B) 30~39 (C) 40~49 (D) 50~59 2 Do
you feel cold? (A) Fear of cold (B) Fear of heat (C) Fear of heat
and cold (D) None of the above 3 Do you have self-hot feeling of
your hand palms? (A) Often (B) Sometimes (C) Seldom (D) Never 4 Do
you easily have hand sweating? (A) Often (B) Sometimes (C) Seldom
(D) Never 5 Do you easily sweat in the day time? (A) Often (B)
Sometimes (C) Seldom (D) Never 6 Do you often feel thirsty and want
to drink? (A) Often (B) Sometimes (C) Seldom (D) Never 7 Do you
feel tired and sleepy? (A) Often (B) Sometimes (C) Seldom (D) Never
8 How is your sleep quality? (A) Good (B) Balance (C) Poor (less
than 5 hours or a lot of dreams) 9 What flavors do you prefer? (A)
Heavy flavor (Spicy) (B) Bland flavor (C) No preference 10 Do you
feel bitter in your mouth? (A) Often (B) Sometimes (C) Seldom (D)
Never 11 Do you have night sweats while you are sleeping? (A) Often
(B) Sometimes (C) Seldom (D) Never 12 What characteristic best
describes you? (A) Impatient (B) Patient (C) Between (A) and (B) 13
Do you suffer from internal heat (e.g., hot flush, oral (A) Often
ulcers)? (B) Sometimes (C) Seldom (D) Never 14 How would you
describe your facial complexion? (A) Very Radiant (B) Radiant (C)
Somewhat Radiant (D) Dull (E) Very Dull 15 What is the color of
your complexion? (A) Red (B) Rosy (C) Below Rosy (D) Pale (E)
Sallow 16 What is your skin condition? (A) Dry skin (B) Normal to
dry (C) Normal or balance skin (D) Normal to oil skin (E) Oily skin
17 What kind of drink do you prefer despite the season? (A) Cold
(B) Hot (C) No preference 18 Do you have acne breakouts on your
face? (A) Yes (B) No 19 Do you have large pores on your face? (A)
Yes (B) No 20 Do you have uneven skin tone? (A) Yes (B) No 21 Do
you have pigmentation issues on your face? (A) Yes (B) No 22 Do you
have (facial) dry patches? (A) Yes (B) No 23 Do you have severe dry
(facial) skin? (A) Yes (B) No 24 Do you have facial redness? (A)
Yes (B) No 25 Do you have more lines/wrinkles than peers? (A) Yes
(B) No 26 Do you suffer from overall darkening of skin tone, (A)
Yes darkening of complexion? (B) No 27 Do you have sensitive skin?
(A) Yes (B) No 28 Do you have dark spot? (A) Yes (B) No 29 Do you
have more sagging problem than your peers? (A) Yes Do you typically
experience any of the following (B) No facial skin conditions
during the summer and/or Please check all that applies to you:
winter months? Summer Months Winter Months Itch skin .quadrature.
.quadrature. Scaling/Flaking .quadrature. .quadrature. 30 Dryness
.quadrature. .quadrature. Cracking .quadrature. .quadrature.
Excessive oil .quadrature. .quadrature. Acne breakouts .quadrature.
.quadrature. None of the above .quadrature. .quadrature. 31 Where
on your body do you experience acne (A) Face breakouts, if at all?
(B) Neck (C) Chest (D) Back (E) Do not experience acne breakouts 32
How frequently do you usually experience acne (A) Once a week or
more often breakouts on your face, if at all? (B) About once every
two weeks (C) At least once a month (D) Less than once a month (E)
Never 33 How could you rate your physical health compared (A)
Excellent to others who are your chronological age? (B) Very good
(C) Good (D) Fair E. Poor 34 Do you have difficulty falling asleep?
(A) Yes (B) No 35 When you have difficulty staying asleep, for what
(A) Stress reasons is your sleep disturbed? (B) Snoring (C)
Physical discomfort (D) Room too warm (E) Noise (F) Light (G)
Smells (H) Other: (I) Don't know 36 Are you usually very tired and
sleepy upon waking (A) Yes in the morning and want to sleep more?
(B) No 37 Do you tend to get very sleepy in the evening before (A)
Yes bedtime? (B) No 38 Do you have seasonal allergies or have
fever? (A) Yes (B) No 39 What is the severity of your allergies?
(A) Very mild (B) Mild (C) Moderate (D) Severe 40 How frequently do
you take vitamin supplements? (A) Everyday (B) 5-6 days a week (C)
3-4 days a week (D) 1-2 days a week (E) Less often than once a week
(F) Never 41 How many days a week do you exercise? (A) Everyday (B)
5-6 days a week (C) 3-4 days a week (D) 1-2 days a week (E) Less
often than once a week (F) Never 42 How often do you skip meals?
(A) Never (B) Sometimes (C) Often (D) Always 43 Which of the
following best describe your typical (A) Well balanced meal meal?
(B) Somewhat balanced meal (C) Not a balanced meal 44 On average,
how many caffeinated drinks do you (A) More than 4 cups/drinks have
everyday? By caffeinated drinks, we mean (B) 3-4 cups/drinks
caffeine energy drinks, coffee, tea and soda. (C) Less than 2
cups/drinks (D) None 45 Thinking about your overall eating habits,
how (A) Extremely healthy healthy do you eat? (B) Very healthy (C)
Somewhat healthy (D) A little healthy (E) Not at all healthy 46 How
would you rate your emotional well-being? (A) Excellent (B) Very
good (C) Good (D) Fair (E) Poor 47 Thinking about your life, in
general, how stressful (A) Extremely stressful would you say your
life is? (B) Very stressful (C) Somewhat stressful (D) Slightly
stressful (E) Not at all stressful 48 How well do you feel you deal
with stress (A) Extremely well (B) Very well (C) Somewhat well (D)
Slightly well (E) Not at all well
[0061] After filling out the questions on the questionnaire, each
participant was directed to a station with a trained skin
specialist, who evaluated the following skin conditions of the
participant based on a scale of 0 to 10: [0062] (1) Acne breakout;
[0063] (2) Pores; [0064] (3) Skin tone; [0065] (4) Age spots;
[0066] (5) Redness; [0067] (6) Global facial wrinkles; [0068] (7)
Crow's feet; [0069] (8) Moist level; [0070] (9) Elasticity; [0071]
(10) Dull-Radiance; and [0072] (11) Skin color variants.
[0073] Standard photo scales were developed for the specialist
evaluation purposes, based on empirical observations of skin
conditions of the Asian sub-population from which the 133
participants were sampled.
[0074] After the evaluation by the skin specialist, each
participant was directed to a TCM doctor, who asked the participant
various questions typical for TCM diagnosis purposes and observed
various physiological characteristics of the participant (e.g.,
pulse, tongue color, tongue fur color, facial complexion, and the
like) and then rendered a diagnosis for the participant with
respect to the skin composition of such participant according to
TCM principles. Specifically, the diagnosis rendered by the TCM
doctor categorized each participant into one of five (5) skin
composition categories, i.e., Yang-dominant, balanced-to-Yang,
balanced, balanced-to-Yin, and Yin-dominant.
[0075] The participants were also subject to testing by various
devices and instruments, which measured various physiological
parameters of the participants, such as skin surface hydration,
sebum level, skin elasticity, skin pH value, the amount of melanin
contained in the skin cells, erythema, skin temperature,
transdermal water-loss (TEWL), skin redness, and radiance-dullness
of the skin.
[0076] A step-wise, multi-class linear discriminant analysis was
then carried out to analyze the testing data as collected
hereinabove. Based on the TCM's doctor's diagnosis, the 133
participants were categorized into 5 groups, i.e., Yang-dominant,
balanced-to-Yang, balanced, balanced-to-Yin, and Yin-dominant. The
answers provided by the participants to the 48 questions on the
questionnaire and the 0-10 scaled evaluations of the participants'
skin conditions as provided by the trained skin specialist were all
treated as independent variables or predictors (V), each of which
was representative of one type of biological and/or psychological
parameter of the participants for subsequent analysis. The
medium/mean values and distributions of the independent variables
(V) for each group of participants of a particular skin composition
were separately calculated.
[0077] An F ratio was then calculated for each one of these
independent variables (V) as an indicator of the capability of each
particular independent variable (V) in discriminating between
different groups of participants of various skin compositions as
determined by the TCM practitioner. The independent variable (V)
with the largest F ratio was initially selected and taken out of
the pool of independent variables (V). F ratios for the remaining
independent variables (the pool excluding the already selected
independent variable) were then recalculated, and the independent
variable with the largest F ratio among these was selected and
taken out of the pool of independent variables. The calculation and
selection steps were reiterated to select multiple independent
variables, which were then used to construct classification
functions for determining which category of skin composition a new
user most likely belongs. Because there were 5 categories of skin
compositions, i.e., Yang-dominant, balanced-to-Yang, balanced,
balanced-to-Yin, and Yin-dominant, five (5) different
classification functions were constructed, each of which
corresponded with a specific skin composition category and each was
used to compute a classification score (Y) for a new user by
entering the biological and/or psychological information of such
new user.
[0078] Specifically, the five (5) different classification
functions so constructed were as follows:
Y.sub.A=C.sub.A+W.sub.A1.times.V.sub.1+W.sub.A2.times.V.sub.2+W.sub.A3.t-
imes.V.sub.3+ . . .
W.sub.A52.times.V.sub.52+W.sub.A53.times.V.sub.53, (1)
Y.sub.B=C.sub.B+W.sub.B1.times.V.sub.1+W.sub.B2.times.V.sub.2+W.sub.B3.t-
imes.V.sub.3+ . . .
W.sub.B52.times.V.sub.52+W.sub.B52.times.V.sub.53, (2)
Y.sub.C=C.sub.C+W.sub.C1.times.V.sub.1+W.sub.C2.times.V.sub.2+W.sub.C3.t-
imes.V.sub.3+ . . .
W.sub.C52.times.V.sub.52+W.sub.C53.times.V.sub.53, (3)
Y.sub.D=C.sub.D+W.sub.D1.times.V.sub.1+W.sub.D2.times.V.sub.2+W.sub.D3.t-
imes.V.sub.3+ . . .
W.sub.D52.times.V.sub.52+W.sub.D53.times.V.sub.53, and (4)
Y.sub.E=C.sub.E+W.sub.E1.times.V.sub.1+W.sub.E2.times.V.sub.2+W.sub.E3.t-
imes.V.sub.3+ . . .
W.sub.E52.times.V.sub.52+W.sub.E53.times.V.sub.53, (5)
[0079] wherein variables V.sub.1 through V.sub.53 are the variables
selected through the discriminant analysis (note that there some of
the answers corresponding to the respondent questionnaire were
broken down into two or more variables, so the total number of
variables was larger than that of the questions); wherein Y.sub.A,
Y.sub.B, Y.sub.C, Y.sub.D and Y.sub.E are the classification scores
calculated for a new user indicative of the likelihood of such user
to be categorized as Yin-dominant (hereinafter referred to as "A"),
balanced-to-Yin (hereinafter referred to as "B"), balanced
(hereinafter referred to as "C"), balanced-to-Yang (hereinafter
referred to as "D"), and Yang-dominant (hereinafter referred to as
"E") skin composition, respectively; wherein W.sub.A1 to W.sub.A53
are coefficients that each correspond with one of the selected
independent variables V.sub.1 through V.sub.53 for skin composition
A (i.e., Yin-dominant) and are indicative of the relative weights
of the selected variables in computing the classification score
Y.sub.A for using the first classification function (1) in
computing the likelihood of such user to be categorized as having
skin composition A (i.e., Yin-dominant); wherein W.sub.B1 to
W.sub.B53 are coefficients that each correspond with one of the
selected independent variables V.sub.1 through V.sub.53 for skin
composition B (i.e., balanced-to-Yin) and are indicative of the
relative weights of the selected variables in computing the
classification score Y.sub.B for using the second classification
function (2) in computing the likelihood of such user to be
categorized as having skin composition B (i.e., balanced-to-Yin);
wherein W.sub.C1 to W.sub.C53 are coefficients that each correspond
with one of the selected independent variables V.sub.1 through
V.sub.53 for skin composition C (i.e., balanced) and are indicative
of the relative weights of the selected variables in computing the
classification score Y.sub.C for using the third classification
function (3) in computing the likelihood of such user to be
categorized as having skin composition C (i.e., balanced); wherein
W.sub.D1 to W.sub.D53 are coefficients that each correspond with
one of the selected independent variables V.sub.1 through V.sub.53
for skin composition D (i.e., balanced-to-Yang) and are indicative
of the relative weights of the selected variables in computing the
classification score Y.sub.D for using the fourth classification
function (4) in computing the likelihood of such user to be
categorized as having skin composition D (i.e., balanced-to-Yang);
wherein W.sub.E1 to W.sub.E53 are coefficients that each correspond
with one of the selected independent variables V.sub.1 through
V.sub.53 for skin composition E (i.e., Yang-dominant) and are
indicative of the relative weights of the selected variables in
computing the classification score Y.sub.E for using the fifth
classification function (5) in computing the likelihood of such
user to be categorized as having skin composition E (i.e.,
Yang-dominant); wherein C.sub.A, C.sub.B, C.sub.C, C.sub.D and
C.sub.E are constants each associated with a specific skin
composition A, B, C, D, and E, respectively.
[0080] The following tables identify each independent variable
V.sub.1 through V.sub.53 and the corresponding coefficients and
constants associated therewith for each classification
function:
TABLE-US-00002 TABLE 2 SELECTED VARIBLES No. Variables V.sub.1
Specialist-Q1-ACNE BREAKOUT (0 TO 10 SCALE) V.sub.2
Specialist-Q2-PORES (0 TO 10 SCALE) V.sub.3 Specialist-Q3-SKIN TONE
(0 TO 10 SCALE) V.sub.4 Specialist-Q5-REDNESS (0 TO 10 SCALE)
V.sub.5 Specialist-Q6-GLOBAL FACIAL WRINKLES (0 TO 10 SCALE )
V.sub.6 Specialist-Q8-MOIST LEVEL-MOIST (YES = 1, NO = 0) V.sub.7
Specialist-Q9-ELASTICITY (0 TO 10 SCALE) V.sub.8
Specialist-Q10-DULL-RADIANCE (0 TO 10 SCALE) V.sub.9
Specialist-Q11-SKIN COLOR VARIANTS-ROSY (YES = 1, NO = 0) V.sub.10
Respondent-Q2-FEAR OF-COLD (YES = 1, NO = 0) V.sub.11
Respondent-Q2-FEAR OF-HEAT (YES = 1, NO = 0) V.sub.12
Respondent-Q2-FEAR OF BOTH-HEAT AND COLD (YES = 1, NO = 0) V.sub.13
Respondent-Q7-DO YOU FEEL TIRED AND SLEEPY (OFTEN = 4) V.sub.14
Respondent-Q12-CHARACTERISTIC DESCRIBING YOU-IMPATIENT (YES = 1, NO
= 0) V.sub.15 Respondent -Q12-CHARACTERISTIC DESCRIBING YOU-PATIENT
(YES = 1, NO = 0) V.sub.16 Respondent-Q13-SUFFER FROM INTERNAL HEAT
(OFTEN = 4) V.sub.17 Respondent-Q14-DESCRIPTION OF FACIAL
COMPLEXION-(VERY DULL = 5) V.sub.18 Respondent-Q15-COLOR OF
COMPLEXION-ROSY (YES = 1, NO = 0) V.sub.19 Respondent-Q15-COLOR OF
COMPLEXION-SLIGHTLY ROSY (YES = 1, NO = 0) V.sub.20
Respondent-Q15-COLOR OF COMPLEXION-PALE (YES = 1, NO = 0) V.sub.21
Respondent-Q15-COLOR OF COMPLEXION-SALLOW (YES = 1, NO = 0)
V.sub.22 Respondent-Q16-SKIN CONDITION-NORMAL TO DRY (YES = 1, NO =
0) V.sub.23 Respondent-Q16-SKIN CONDITION-NORMAL OR BALANCED (YES =
1, NO = 0) V.sub.24 Respondent-Q16-SKIN CONDITION-NORMAL TO OILY
(YES = 1, NO = 0) V.sub.25 Respondent-Q18-SKIN PROBLEMS-ACNE
BREAKOUTS (YES = 1, NO = 0) V.sub.26 Respondent-Q19-SKIN
PROBLEMS-LARGE PORES (YES = 1, NO = 0) V.sub.27 Respondent-Q24-SKIN
PROBLEMS-FACIAL REDNESS (YES = 1, NO = 0) V.sub.28
Respondent-Q30-FACIAL SKIN CONDITIONS IN SUMMER-ITCHY SKIN (YES =
1, NO = 0) V.sub.29 Respondent-Q30-FACIAL SKIN CONDITIONS IN
SUMMER- SCALING/FLAKING (YES = 1, NO = 0) V.sub.30
Respondent-Q30-FACIAL SKIN CONDITIONS IN SUMMER-DRYNESS (YES = 1,
NO = 0) V.sub.31 Respondent-Q30-FACIAL SKIN CONDITIONS IN
SUMMER-EXCESSIVE OIL (YES = 1, NO = 0) V.sub.32
Respondent-Q30-FACIAL SKIN CONDITIONS IN SUMMER-ACNE BREAKOUTS (YES
= 1, NO = 0) V.sub.33 Respondent-Q30-FACIAL SKIN CONDITIONS IN
WINTER-ITCHY SKIN (YES = 1, NO = 0) V.sub.34 Respondent-Q30-FACIAL
SKIN CONDITIONS IN WINTER- SCALING/FLAKING (YES = 1, NO = 0)
V.sub.35 Respondent-Q30-FACIAL SKIN CONDITIONS IN WINTER-DRYNESS
(YES = 1, NO = 0) V.sub.36 Respondent-Q30-FACIAL SKIN CONDITIONS IN
WINTER-EXCESSIVE OIL (YES = 1, NO = 0) V.sub.37
Respondent-Q30-FACIAL SKIN CONDITIONS IN WINTER-ACNE BREAKOUTS (YES
= 1, NO = 0) V.sub.38 Respondent-Q31-WHERE ON BODY EXPERIENCE ACNE
BREAKOUT- NECK (YES = 1, NO = 0) V.sub.39 Respondent-Q31-WHERE ON
BODY EXPERIENCE ACNE BREAKOUT- CHEST (YES = 1, NO = 0) V.sub.40
Respondent-Q31-WHERE ON BODY EXPERIENCE ACNE BREAKOUT-BACK (YES =
1, NO = 0) V.sub.41 Respondent-Q31-WHERE ON BODY EXPERIENCE ACNE
BREAKOUT-DO NOT EXPERIENCE BREAKOUTS (YES = 1, NO = 0) V.sub.42
Respondent-Q32-FREQUENCY OF ACNE BREAKOUTS ON FACE (5 = ONCE A WEEK
OR MORE) V.sub.43 Respondent-Q35-REASONS FOR SLEEP BEING
DISTURBED-STRESS (YES = 1, NO = 0) V.sub.44 Respondent-Q35-REASONS
FOR SLEEP BEING DISTURBED-SNORING (YES = 1, NO = 0) V.sub.45
Respondent-Q35-REASONS FOR SLEEP BEING DISTURBED-PHYSICAL
DISCOMFORT (YES = 1, NO = 0) V.sub.46 Respondent-Q35-REASONS FOR
SLEEP BEING DISTURBED-ROOM TOO WARM/COLD (YES = 1, NO = 0) V.sub.47
Respondent-Q35-REASONS FOR SLEEP BEING DISTURBED-NOISE (YES = 1, NO
= 0) V.sub.48 Respondent-Q35-REASONS FOR SLEEP BEING
DISTURBED-LIGHT (YES = 1, NO = 0) V.sub.49 Respondent-Q35-REASONS
FOR SLEEP BEING DISTURBED-SMELLS (YES = 1, NO = 0) V.sub.50
Respondent-Q40-FREQUENCY TAKING VITAMIN SUPPLEMENTS (6 = EVERDAY)
V.sub.51 Respondent-Q44-NUMBER OF CAFFEINATED DRINKS CONSUMED EACH
DAY (4 = MORE THAN 4 CUPS) V.sub.52 Respondent-Q45-HOW HEALTHY DO
YOU EAT (5 = EXTREMELY HEALTHY) V.sub.53 Respondent-Q47-HOW
STRESSFUL IS YOUR LIFE (5 = EXTREMELY STRESSFUL)
TABLE-US-00003 TABLE 3 COEFFICIENTS AND CONSTANTS A B C D E
W.sub.A1 -5.03481 W.sub.B1 -3.32657 W.sub.C1 -6.41283 W.sub.D1
-1.9017 W.sub.E1 -1.84041 W.sub.A2 2.86515 W.sub.B2 1.16363
W.sub.C2 -1.39053 W.sub.D2 2.1262 W.sub.E2 4.40778 W.sub.A3 7.25343
W.sub.B3 4.91899 W.sub.C3 4.70919 W.sub.D3 4.97844 W.sub.E3 6.45077
W.sub.A4 0.77215 W.sub.B4 0.3698 W.sub.C4 0.44631 W.sub.D4 2.18927
W.sub.E4 3.21603 W.sub.A5 -13.53785 W.sub.B5 -9.49917 W.sub.C5
-11.61919 W.sub.D5 -14.47177 W.sub.E5 -24.54775 W.sub.A6 -6.26384
W.sub.B6 -4.73265 W.sub.C6 -1.0894 W.sub.D6 -4.09167 W.sub.E6
-14.96101 W.sub.A7 4.91627 W.sub.B7 2.28944 W.sub.C7 1.78096
W.sub.D7 5.3033 W.sub.E7 10.86091 W.sub.A8 9.60978 W.sub.B8
11.04179 W.sub.C8 12.78336 W.sub.D8 14.26397 W.sub.E8 22.76282
W.sub.A9 4.50266 W.sub.B9 4.24891 W.sub.C9 6.78872 W.sub.D9 2.71058
W.sub.E9 1.27952 W.sub.A10 5.79872 W.sub.B10 6.70748 W.sub.C10
6.90218 W.sub.D10 8.06332 W.sub.E10 6.77549 W.sub.A11 2.05002
W.sub.B11 2.81332 W.sub.C11 6.43626 W.sub.D11 7.87991 W.sub.E11
12.42265 W.sub.A12 7.13019 W.sub.B12 7.78881 W.sub.C12 7.78571
W.sub.D12 9.85302 W.sub.E12 7.37807 W.sub.A13 13.0662 W.sub.B13
10.91798 W.sub.C13 12.38146 W.sub.D13 10.65906 W.sub.E13 13.75115
W.sub.A14 5.84626 W.sub.B14 5.01957 W.sub.C14 2.84629 W.sub.D14
3.30128 W.sub.E14 -4.00897 W.sub.A15 14.35454 W.sub.B15 13.74852
W.sub.C15 12.02993 W.sub.D15 12.9599 W.sub.E15 15.45324 W.sub.A16
3.16837 W.sub.B16 3.51648 W.sub.C16 3.20893 W.sub.D16 1.03277
W.sub.E16 -0.70696 W.sub.A17 10.70804 W.sub.B17 10.83141 W.sub.C17
10.46987 W.sub.D17 9.37636 W.sub.E17 8.0314 W.sub.A18 78.28738
W.sub.B18 78.51756 W.sub.C18 79.8523 W.sub.D18 83.91855 W.sub.E18
93.36256 W.sub.A19 81.92217 W.sub.B19 87.84184 W.sub.C19 87.58114
W.sub.D19 87.2457 W.sub.E19 101.40503 W.sub.A20 78.63351 W.sub.B20
86.28866 W.sub.C20 86.93638 W.sub.D20 86.65417 W.sub.E20 102.20674
W.sub.A21 63.50097 W.sub.B21 68.59361 W.sub.C21 70.52393 W.sub.D21
69.55426 W.sub.E21 77.66964 W.sub.A22 3.44182 W.sub.B22 0.15277
W.sub.C22 5.69318 W.sub.D22 -1.12491 W.sub.E22 -4.40742 W.sub.A23
10.0591 W.sub.B23 9.82875 W.sub.C23 12.1122 W.sub.D23 7.59917
W.sub.E23 11.69212 W.sub.A24 11.71474 W.sub.B24 9.46744 W.sub.C24
16.06216 W.sub.D24 9.41853 W.sub.E24 8.32798 W.sub.A25 -2.40053
W.sub.B25 -0.5321 W.sub.C25 -0.40143 W.sub.D25 -0.08659 W.sub.E25
1.20704 W.sub.A26 3.42408 W.sub.B26 2.81038 W.sub.C26 5.34481
W.sub.D26 5.39178 W.sub.E26 9.99927 W.sub.A27 1.58022 W.sub.B27
6.32234 W.sub.C27 3.96203 W.sub.D27 3.04199 W.sub.E27 -0.55476
W.sub.A28 8.87747 W.sub.B28 6.62846 W.sub.C28 1.36545 W.sub.D28
1.69279 W.sub.E28 -1.59018 W.sub.A29 14.4482 W.sub.B29 15.59435
W.sub.C29 7.56485 W.sub.D29 15.52505 W.sub.E29 20.52038 W.sub.A30
-6.32975 W.sub.B30 -9.09096 W.sub.C30 -5.6209 W.sub.D30 -8.26135
W.sub.E30 -10.52059 W.sub.A31 -2.84589 W.sub.B31 -4.68028 W.sub.C31
-6.59683 W.sub.D31 -2.50031 W.sub.E31 -1.83753 W.sub.A32 -6.29274
W.sub.B32 -5.38067 W.sub.C32 -1.85282 W.sub.D32 -5.20866 W.sub.E32
-8.16089 W.sub.A33 0.86533 W.sub.B33 0.21689 W.sub.C33 0.59462
W.sub.D33 -0.1715 W.sub.E33 1.46984 W.sub.A34 -1.00161 W.sub.B34
2.88752 W.sub.C34 3.17145 W.sub.D34 1.02958 W.sub.E34 2.17715
W.sub.A35 17.05898 W.sub.B35 17.2045 W.sub.C35 11.05842 W.sub.D35
16.88556 W.sub.E35 19.40689 W.sub.A36 -10.75367 W.sub.B36 -8.30602
W.sub.C36 -8.61707 W.sub.D36 -5.04753 W.sub.E36 -8.40332 W.sub.A37
0.36254 W.sub.B37 -3.49799 W.sub.C37 -2.80702 W.sub.D37 -1.18985
W.sub.E37 -1.14465 W.sub.A38 -19.32756 W.sub.B38 -15.42999
W.sub.C38 -16.92841 W.sub.D38 -20.13988 W.sub.E38 -32.79576
W.sub.A39 0.33108 W.sub.B39 0.28936 W.sub.C39 6.0703 W.sub.D39
3.56482 W.sub.E39 8.39871 W.sub.A40 -0.50417 W.sub.B40 -1.56768
W.sub.C40 -1.52976 W.sub.D40 -1.44652 W.sub.E40 -1.43144 W.sub.A41
40.07422 W.sub.B41 34.64281 W.sub.C41 34.16851 W.sub.D41 38.45052
W.sub.E41 45.4052 W.sub.A42 12.50723 W.sub.B42 12.69321 W.sub.C42
12.30448 W.sub.D42 13.63575 W.sub.E42 15.64608 W.sub.A43 -9.17133
W.sub.B43 -6.97504 W.sub.C43 -1.10747 W.sub.D43 -3.99002 W.sub.E43
-2.35332 W.sub.A44 0.24349 W.sub.B44 -0.63206 W.sub.C44 -0.78648
W.sub.D44 0.14629 W.sub.E44 3.93916 W.sub.A45 -8.63626 W.sub.B45
-5.86896 W.sub.C45 -7.67355 W.sub.D45 -4.62572 W.sub.E45 -8.14927
W.sub.A46 -3.71256 W.sub.B46 -3.49322 W.sub.C46 -1.32212 W.sub.D46
-1.31128 W.sub.E46 2.40936 W.sub.A47 12.48024 W.sub.B47 10.77945
W.sub.C47 10.59542 W.sub.D47 11.26331 W.sub.E47 14.35998 W.sub.A48
-8.06078 W.sub.B48 -6.94508 W.sub.C48 -7.48413 W.sub.D48 -8.41228
W.sub.E48 -10.89619 W.sub.A49 -18.75243 W.sub.B49 -18.41954
W.sub.C49 -17.46166 W.sub.D49 -17.98823 W.sub.E49 -22.52914
W.sub.A50 1.51757 W.sub.B50 2.04889 W.sub.C50 1.62452 W.sub.D50
1.24405 W.sub.E50 0.77243 W.sub.A51 6.04014 W.sub.B51 6.03027
W.sub.C51 4.3963 W.sub.D51 5.5696 W.sub.E51 6.28266 W.sub.A52
2.80341 W.sub.B52 3.74745 W.sub.C52 3.70402 W.sub.D52 3.80916
W.sub.E52 4.06269 W.sub.A53 0.4751 W.sub.B53 -1.04686 W.sub.C53
-1.96563 W.sub.D53 0.80047 W.sub.E53 2.49099 C.sub.A -151.32817
C.sub.B -143.07719 C.sub.C -144.5403 C.sub.D -155.94759 C.sub.E
-240.69533
[0081] The classification score Y.sub.A, Y.sub.B, Y.sub.C, Y.sub.D
and Y.sub.E outputted by each classification function represents
the likelihood of a new user being classified by a TCM practitioner
as having the specific skin composition A, B, C, D and E associated
with the respective classification function. In other words, for
each new user, each classification function will compute a
classification score, and the skin composition associated with the
classification function that renders the highest classification
score is the skin composition of the user.
[0082] While the invention has been described in connection with a
preferred embodiment, it is not intended to limit the scope of the
invention to the particular form set forth but, on the contrary, it
is intended to cover such alternatives, modifications, and
equivalents as may be included within the spirit and scope of the
invention as defined by the appended claims.
* * * * *